JP2004152205A - Forecast value evaluation method, system and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To find a difference in perspectives and facilitate determination of a cause for the difference in perspectives and correction of forecast values even when the forecast values formed based on different factors respectively by two companies in business accidentally match each other in a joint forecast system for forming a forecast jointly by the two sharing companies. <P>SOLUTION: Two companies share both a forecast value and basis information obtained by patternizing a factor for calculating the forecast value. When such conditions as those under which the basis information of the two companies does not match each other are satisfied, it is determined that there is a difference in perspectives of the forecast value. When it is so determined, then the basis information is presented to the two companies, a recommendable value for correcting the forecast value is calculated based on the basis information, and the calculated value is presented to the two companies. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a commodity demand forecasting method, and more particularly, to a method in which an orderer and a supplier jointly predict demand through a network represented by the Internet.
[0002]
[Prior art]
2. Description of the Related Art In recent years, supply chain management has become popular in which stock information and sales results are shared across companies, thereby streamlining the process from material procurement to product delivery and reducing inventory. Among them, by sharing forecasts of sales volume and the like between companies, CPFR (Collaborative Planning, Forecasting) that optimizes production plans and order plans, reduces out-of-stock rates, reduces inventory, improves sales, etc. and Replenishment) is gaining attention. CPFR is defined by VICS (Voluntary Interindustry Commander Standards).
[0003]
In the conventional CPFR system, first, a sales company that sells a product and a purchasing company that purchases the product create a predicted value that is a prediction of a sales amount, an order amount, and the like, and register the predicted value in a prediction sharing server. Here, the sales amount is the amount of goods sold by the purchasing company to other companies or individuals, and the order amount is the amount of goods ordered by the purchasing company to the sales company.
[0004]
Next, when the predicted value matches the exception condition for finding an inappropriate predicted value, the prediction sharing server determines that the exception is an exception, and notifies the selling company and the purchasing company. As the exceptional condition, a condition is used in which the predicted values of the registered sales company and the purchase company are compared, and the difference or ratio exceeds a threshold value set in advance by the sales company and the purchase company. The notified sales company and purchasing company consult with each other to determine the cause of the exception and correct the predicted value.
[0005]
[Patent Document 1]
JP-A-5-173602
[0006]
[Problems to be solved by the invention]
However, in the conventional CPFR system, the exceptional condition is applied only to the predicted value and the exception is determined, so that there are the following problems.
[0007]
First, a sales company and a purchasing company cannot judge that a predicted value created based on different factors accidentally does not match an exceptional condition. For example, for a product with an average sales volume of 100 units, the purchasing company sells 50% off, so the sales volume is predicted to be 150%, which is 50% higher than usual. Is predicted to be 150, which is 50% more than usual. In fact, the synergistic effect of a 50% discount sale and a magazine feature could result in a sales volume of more than 50%, which should be notified to both companies as an exception. Since it does not exceed the threshold value, it cannot be determined as an exception.
[0008]
Secondly, if the predicted values of the sales company and the purchasing company are determined to be exceptions, the cause of the exception is not known, so the forecasting person of the sales company and the prediction person of the purchasing company consult with each exception to find out the cause. Must. For example, with respect to a product having an average sales volume of 100, the purchasing company conducts a 50% discount sale, so that the sales volume is predicted to be 50 units, which is 50% higher than usual, and the sales company is predicted to be 100 units as usual. I do. In this case, although exceptions occur, both the selling company and the purchasing company can refer to only the predicted values of 150 and 100, and the basis for calculating the predicted values is unknown. Otherwise, the prediction cannot be corrected.
[0009]
Third, the predicted value must be recalculated and corrected based on the determined cause. In the case of the above example, it is assumed that the selling company consults the purchasing company and finds out that the exception is that the selling company did not recognize the 50% discount sale. In this case, the forecast value must be calculated and corrected again by the in-house system or the person in charge in consideration of the 50% discount sale.
[0010]
SUMMARY OF THE INVENTION An object of the present invention is to provide an exception determination method capable of finding a difference in views on future demand that cannot be found because predicted values coincide with each other.
[0011]
Another object of the present invention is to facilitate identification of an exception factor when it is determined that an exception has occurred.
[0012]
Another object of the present invention is to facilitate the correction of predicted values.
[0013]
[Means for Solving the Problems]
In order to achieve the above object, in a system in which a terminal of a sales company or a terminal of a purchasing company and a prediction sharing server are connected via the Internet, in addition to the predicted value, a basis for patterning a factor for calculating the predicted value. Information is also shared, an exception is determined in the following procedure, and a predicted value and a recommended value for correcting the basis information are calculated.
[0014]
First, the sales company and the purchasing company jointly define the basis information and the importance of the significance of each basis information in numerical form, and register them in the prediction sharing server. For example, “sale 50% OFF” and “magazine feature” are defined as the basis information.
[0015]
Also, a predicted value exception condition for determining an exception from the predicted value of the sales company and the purchasing company, and a ground information exception condition for determining the exception from the ground information of the sales company and the purchased company, when an exception is determined. A selling company and a purchasing company respectively register in a prediction sharing server a recommended predicted value calculating method for calculating a recommended value of a predicted value and a recommended ground information calculating method for calculating a corrected value of ground information. For example, “predicted value of sales company−predicted value of purchased company <threshold value” is defined as the predicted value exception condition. For example, the basis information exception condition defines “basis information of the selling company 根 basis information of the purchasing company”. The recommended prediction value calculation method defines, for example, "match the prediction value of the selling company to the prediction value of the purchasing company". The recommended basis information calculation method defines, for example, "Match the basis information of the selling company to the basis information of the purchasing company".
[0016]
Next, the selling company and the purchasing company register the created prediction values and the basis information in the prediction sharing server.
[0017]
First, the prediction sharing server determines whether the predicted values of the selling company and the purchasing company apply to the exceptional condition of the predicted value, and if so, determines that the exception is present. Also, it is determined whether the basis information of the selling company and the purchasing company applies to the basis information exceptional condition, and if so, an exception is determined. At the time of the exception determination of the ground information, it is also possible to specify the ground information to be used for the determination according to the importance, for example, to make an exception determination only with the ground information having a specific importance or higher.
[0018]
When it is determined that either the predicted value or the ground information is an exception, a recommended predicted value and recommended ground information are calculated according to the following procedure, and the user is notified.
[0019]
First, a recommended predicted value is calculated from predicted values of a sales company and a purchasing company based on a recommended predicted value calculation method. Next, the recommended basis information is calculated from the basis information of the selling company and the purchasing company based on the recommended basis information calculation method. When calculating the recommended ground information, it is also possible to specify the ground information to be used for the calculation based on the importance, for example, to calculate the recommended ground information using only ground information having a specific importance or higher.
[0020]
Thereafter, the prediction sharing server notifies the sales company and the purchase company of the exception, the recommended predicted value, and the recommended basis information. Before the notification, the predicted value and the basis information may be automatically corrected based on the recommended predicted value and the recommended basis information.
[0021]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0022]
FIGS. 1, 3, 4, and 12 show the processing flow of this embodiment, FIGS. 2, 5, 10, 11, 14, and 15 show screen images of this embodiment, and FIGS. 9 and 13 show the tables of the present embodiment, respectively.
[0023]
First, an overall configuration diagram of the prediction sharing system shown in FIG. 1 will be described.
[0024]
The prediction sharing system includes a prediction sharing server 11, a sales company terminal 12, a purchasing company terminal 13, and a communication network 14 connecting these terminals. The communication network 14 can use the Internet, a LAN, or any other wired or wireless communication network.
[0025]
The prediction sharing server 11 is an information processing device such as a workstation server installed in a selling company, a purchasing company, or another company. The prediction sharing server 11 receives the prediction value and the basis information, the exceptional condition, the recommended prediction value calculation method, and the recommended basis information calculation method, which will be described later, transmitted from the sales company terminal 12 and the purchase company terminal 13, and based on the received information. In addition, it has a function of performing exception determination and calculating a recommended predicted value and recommended ground information, and correcting the predicted value and the ground information, and transmitting the result to the sales company terminal 12 and the purchase company terminal 13.
[0026]
The sales company terminal 12 is an information processing device such as a personal computer installed in a sales company or the like. The sales company terminal 12 has a function of transmitting a predicted value and basis information calculated by a sales company representative or an in-house system of the sales company to the prediction sharing server 11 and an exception described later determined by a sales company representative or the like. It has a function of transmitting a condition, a method of calculating a recommended predicted value, and a method of calculating recommended ground information to the prediction shared server 11, and a function of displaying exceptions, recommended predicted values, and recommended ground information described below transmitted from the predicted shared server 11.
[0027]
The purchasing company terminal 13 is an information processing device such as a personal computer installed in a purchasing company or the like. The purchasing company terminal 13 has a function of transmitting a predicted value and basis information calculated by a person in charge of the purchasing company or an in-house system of the purchasing company to the prediction sharing server 11, and an exception described later determined by a person in charge of the purchasing company. It has a function of transmitting a condition, a method of calculating a recommended predicted value, and a method of calculating recommended ground information to the prediction shared server 11, and a function of displaying exceptions, recommended predicted values, and recommended ground information described below transmitted from the predicted shared server 11.
[0028]
In the present embodiment, the prediction sharing server 11, the sales company terminal 12, and the purchasing company terminal 13 are configured by three separate information processing devices. However, the prediction sharing server 11 is the same as the sales company terminal 12 or the purchasing company terminal 13. An information processing device may be used.
[0029]
The above-described predicted value is a predicted value of a specific item related to a product sold by a selling company to a purchasing company. Items include a sales forecast, which is the quantity of products sold by the purchasing company to the company and the individual, and an order prediction, which is the quantity of products purchased by the purchasing company from the selling company. The predicted value is created by the selling company and the purchasing company independently for each combination of the product type and the period for the same item.
[0030]
The term “period” refers to any fixed time such as one day, one week, one month, etc. For example, when one week is defined as one period, June 1 week, June 2 weeks, June 3 weeks, etc. Period exists. The selling company and the purchasing company create the predicted value and the basis information to be described later according to the same period definition.
[0031]
The basis information described above is obtained by patterning events such as a bargain sale and an advertisement that affect the predicted value. The basis information to be used is defined in advance by the selling company and the purchasing company, and is stored in the prediction sharing server 11. For the basis information, a sales company and a purchasing company independently create zero or more pieces for each combination of the product type and the period.
[0032]
The exception described above is a state in which the prediction value or the basis information independently created by the selling company and the purchasing company matches the exception condition, and is determined by the prediction sharing server 11. The exceptional condition is created by the selling company and the purchasing company independently or jointly, and is a rule for judging a difference between predicted values and a difference in basis information. For example, a range in which the difference between the predicted values created by the sales company and the purchase company is unacceptable, and a situation in which the basis information does not match correspond. The exception conditions will be described later.
[0033]
The above-mentioned recommended prediction value calculation method and recommended ground information calculation method are methods for calculating a predicted value determined as an exception and a recommended value for correcting the ground information, and the sales company and the purchasing company independently or jointly operate. And registered in the prediction sharing server 11. The prediction sharing server 11 creates a recommended predicted value and recommended basis information from the predicted value and the basis information based on the setting. The recommended prediction value and the recommended basis information are collectively called a recommended value, and the recommended predicted value calculation method and the recommended basis information calculation method are collectively called a recommended value calculation method.
[0034]
The processing flow of the present embodiment shown in FIG. 1 will be described.
[0035]
First, in step S101, the prediction sharing server 11 receives the exception condition and the recommended value calculation method transmitted from the sales company terminal 12 and the purchase company terminal 13, and stores the exception condition and the recommended value calculation method table T81.
[0036]
Next, the prediction sharing server 11 receives the predicted value and the basis information transmitted from the selling company terminal 12 and the purchasing company terminal 13 in step S102, and stores the predicted value and the basis information table T61.
[0037]
In step S103, the prediction sharing server 11 determines an exception based on the information in the exception condition / recommended value calculation method table T81, the ground information definition table T71, and the predicted value / ground information table T61, and recommends the predicted value and the recommended ground information. Is calculated, and the result is stored in the exception / recommended predicted value / recommended basis information table T91.
[0038]
Finally, in step S104, the prediction sharing server 11 notifies the sales company terminal 12 and the purchase company terminal 13 of the exception, the recommended prediction value, and the recommended basis information based on the information of the exception / recommended predicted value / recommended basis information table T91. I do. E-mail or the like is used for the notification method, and the destination address is obtained from the company table T31.
[0039]
Hereinafter, each method will be further described with reference to the drawings.
[0040]
First, a method of registering the exception condition / recommended value calculation method (step S101) will be described.
[0041]
The selling company terminal 12 or the purchasing company terminal 13 displays the exception condition / recommended value calculation method setting screen 21 shown in FIG. 2 on the WWW browser by accessing the prediction sharing server 11. On this screen, one exception condition and one recommended value calculation method can be set by inputting the setting items described below.
[0042]
In the exception condition / recommended value calculation method setting screen 21, the target predicted value / ground information box 22 is used to specify the exception condition set on this screen and the predicted value and the ground information to which the recommended value calculation method is applied. . The creation company field 221 is for setting the creation company number for creating the exception condition / recommended value calculation method, and the transaction company field 222 is for setting the transaction company number of the creation company. The designation of the creation company field 221 and the transaction company field 222 uses a company number 311 of a company table T31 described later. The product number field 223 is for setting a product number for identifying a product type. Fields 224 and 225 are used to set the range of the target period. The field 224 specifies the number of weeks after the exception search processing date to start the period, and the field 225 specifies the exception search processing date. Specify the number of weeks after the end of the period. In FIG. 2, when 1 and 4 are entered, when the exceptional search processing day is May 4th week, the week after June 1 week, which is one week after May 4th week, is a week after 4 weeks. It is set that the forecast value and the basis information up to June 4 are the period of the exception determination. Here, the range of the relative period is specified, but a method of directly specifying the period, such as the first week of June to the fourth week of June, may be used.
[0043]
The predicted value exception condition box 25 is for setting an exception condition for the predicted value. The setting of the exception condition is performed by selecting any one of the radio buttons 251 to 257. When the radio button 251 is selected, the predicted value is not used for the exception determination. When the radio button 257 is selected from the radio buttons 252, an exception condition is determined which is determined to be an exception when the predicted value of the creating company and the predicted value of the trading company apply to the expression described on the right of the radio button. Here, the exception threshold value on the right side of each conditional expression is set in the exception threshold field 258.
[0044]
The target ground information box 23 is for designating ground information to be subjected to exception determination and recommended value calculation. The specification of the basis information is performed by selecting any one of the radio buttons 231 to 234. When the radio button 231 is selected, all the basis information is targeted. When the radio button 232 is selected, only the basis information in which the importance is equal to or higher than the importance set in the minimum importance field 233 is targeted. When the radio button 234 is selected, only a plurality of pieces of basis information having the highest importance are selected from among a plurality of pieces of basis information limited by a target period and a product number at the time of exception determination of the basis information and calculation of a recommended value. Are the targets of exception determination and calculation of recommended basis information. In addition to the target basis information shown here, "the basis information having the nth (n: an arbitrary natural number) th significant importance is included", "the basis of importance smaller than the importance of the most important basis information by n" Even if the target rationale information is specified by "Target the information up to the information", "Target the lowest importance base information", "Target the importance lower than x (x: arbitrary importance)" or the like Good.
[0045]
The ground information exception condition box 26 is for setting an exception condition for the ground information. The setting of the exception condition is performed by selecting the condition of the check box 265 from the check box 261. A plurality of check boxes 261 to 265 can be selected at the same time, and an exceptional condition that is determined as an exception when any of the selected conditions is satisfied is set. If all the check boxes are not selected, the basis information is not used for the exception judgment.
[0046]
The recommended predicted value calculation method box 27 is for setting a method of calculating a recommended predicted value when it is determined that an exception has occurred. The setting of the recommended prediction value is performed by selecting one of the methods from the radio button 271 to the radio button 273. In addition to the methods indicated by the radio buttons 271 to 273, "match the predicted values of both companies to the larger one", "match the predicted values of both companies to the smaller one", and To the average of the predicted values of ".
[0047]
The recommended basis information calculation method box 28 is for setting a method for calculating recommended basis information when it is determined that an exception has occurred. The setting of the recommended basis information is performed by selecting one of the methods of the radio button 281 to the radio button 285.
[0048]
The automatic correction box 24 is for setting whether or not to automatically correct the predicted value and the ground information based on the recommended predicted value and the recommended ground information calculated by the recommended predicted value calculating method and the recommended ground information calculating method. .
[0049]
The notification box 29 is used to set whether or not to transmit the notification screen 41 shown in FIG. 10 to the creating company or the trading company by mail.
[0050]
When the add button 211 is pressed after setting each field of the exception condition / recommended value calculation method setting screen 21, the exception condition and the recommended value calculation method are additionally registered in the exception condition / recommended value calculation method table T81. When the cancel button 212 is pressed, the exception condition / recommended value calculation method is not registered and the exception condition / recommended value calculation method setting screen 21 is closed and the process ends.
[0051]
Here, the exception condition and the recommended value calculation method are registered using the exception condition / recommended value calculation method setting screen 21. However, the contents of these settings are described in a file or the like from the sales company terminal 12 or the purchase company terminal 13 or the like. May be received by the prediction sharing server 11 and registered in the exception condition / recommended value calculation method table T81.
[0052]
Here, a company table T31 for storing information about each company sharing the prediction will be described with reference to FIG. The company table T31 stores information on one company in one line. As information on the company, a company number 311 and a mail address 312 which are identification numbers of the company are stored. Thereafter, the company number 311 defined in the company table T31 is stored in the column for specifying the company in each table.
[0053]
Next, the exception condition / recommended value calculation method table T81 will be described with reference to FIG.
[0054]
The exception condition / recommended value calculation method table T81 stores one exception condition / recommended value calculation method set on the exception condition / recommended value calculation method setting screen 21 in one line.
[0055]
The exception condition number 811 which is the identification number of the exception condition assigned by the prediction sharing server 11 is stored, the creation company field 221 is set to the creation company 812, the business company field 222 is set to the business company 813, and the product number field 223 is set to the product number. At 814, the field 224 is stored at the beginning of the period 815 and the field 225 at the end of the period 816.
[0056]
Of the radio buttons 231 to 234, the number in parentheses described to the right of the selected radio button is used as the target basis information 817, the minimum importance field 233 is used as the minimum importance 818, and the radio button 251 is used as the radio button. Of the buttons 257, the number in parentheses described to the right of the selected radio button is set to the predicted value exception condition 819, the exception threshold field 258 is set to the exception threshold 820, and the check box 261 is selected. “○”, if not selected, “×” is used as the basis information relation 1 (821), and similarly, whether the check box 265 is selected from the check box 262 to the basis information relation 2 (822) is used as the basis information relation. 5 (825), the radio button 271 to the radio button 273 are described to the right of the selected radio button. The numbers in parentheses in the parentheses are used for the recommended prediction value calculation method 826, and the numbers in parentheses described to the right of the selected radio button among the radio buttons 281 to 285 are used for the recommended basis information calculation method 827. If "Yes" is selected in the automatic correction box 24, "O" is selected. If "No" is selected, "X" is set to automatic correction 828. If the check box 291 is selected, "O" is selected. If not, "x" is stored in the creation company notification 829, and similarly, whether or not the check box 292 is selected is stored in the transaction company notification 830.
[0057]
For example, when the setting is made as in the exception condition / recommended value calculation method setting screen 21 shown in FIG. 2 and the add button 211 is pressed, the exception condition / recommended value calculation method table T81 is stored as a row 8101 in the exception condition / recommended value calculation method table T81.
[0058]
Next, the predicted value and the ground information received by the prediction sharing server 11 in the predicted value / ground information receiving process (step S102) will be described.
[0059]
First, the ground information definition table T71 for storing the ground information definition will be described with reference to FIG.
[0060]
The basis information definition table T71 stores one basis information definition per line. As the basis information definition contents, a basis information number 711, which is an identification number of the basis information, a basis information name 712, and a degree of importance 713 are stored. Hereinafter, the basis information number 711 defined in the basis information definition table T71 is stored in the basis information of each table, and when storing a plurality of basis information, the basis information number 711 is also written separated by a comma (,). I do. The basis information 712 is displayed on the basis information of the exception screen 51, the notification screen 41, and the notification screen 45.
[0061]
Next, the predicted value / foundation information table T61 will be described with reference to FIG.
[0062]
The forecast value / basis information table T61 includes a forecast creation company 611 that is a company that created the forecast, a forecast sharing company 612 that is a company sharing the forecast, a product number 613, a period 614, a forecast value 615, and basis information 616. save.
[0063]
For example, the predicted value of the product (product number: 0001) that the selling company (company number: 1001) trades with the purchasing company (company number: 1002) in the first week of July is 196, and the basis information is “sale 20%” ( When it is created as the basis information number: 2) and transmitted from the sales company terminal 12 to the prediction sharing server 11, the prediction sharing server 11 stores the information as shown in a row 6105 in FIG.
[0064]
Next, the exception determination / recommended value calculation (step S103) processing will be described using the flow shown in FIG.
[0065]
In step S301, the prediction sharing server 11 makes an exception determination by a method described later, and writes the exception to the exception / recommended predicted value / recommended basis information table T91. For each exception written in step S301, the exception, the recommended predicted value, and the recommended basis information table T91, the process from step S302 to step S316 is repeated. First, in step S303, the recommended prediction of the exception condition / recommended place calculation method which is the exception condition number 811 of the exception condition / recommended value calculation method table T81 which is the same as the exception condition number 912 of the exception / recommended predicted value / recommended basis information table T91. The value calculation method 826, the recommended basis information calculation method 827, and the automatic correction 828 are read.
[0066]
In step S304, a recommended predicted value is calculated based on the recommended predicted value calculation method 826 read in step S303. It is determined in step S305 whether or not the recommended predicted value of the creating company 812 has been calculated. If calculated, the process proceeds to step S306, and if not, the process proceeds to step S307. In step S306, the created company recommended predicted value is written in the exception / recommended predicted value / recommended basis information table T91, and the process proceeds to step S307. In step S307, it is determined in step S304 whether or not the recommended predicted value of the trading company 813 has been calculated. If the recommended predicted value has been calculated, the process proceeds to step S308. If not, the process proceeds to step S309. In step S308, the transaction company recommended predicted value is written in the exception / recommended predicted value / recommended basis information table T91, and after completion, the process proceeds to step S309.
[0067]
In step S309, recommended basis information is calculated based on the recommended basis information calculation method 827. It is determined in step S310 whether recommended basis information of the creating company 812 has been calculated, and if calculated, the process proceeds to step S311. If not, the process proceeds to step S312. In step S311, the created company recommended basis information is written in the exception / recommended predicted value / recommended basis information table T91, and after completion, the process proceeds to step S312. In step S312, it is determined in step S309 whether or not the recommended basis information of the trading company has been calculated. If it has been calculated, the process proceeds to step S313, and if not, the process proceeds to step S314. In step S313, the transaction company recommended basis information is written in the exception / recommended predicted value / recommended basis information table T91, and after completion, the process proceeds to step S314.
[0068]
In step S314, based on the automatic correction 828 read in step S303, the process proceeds to step S316 if the automatic correction is not to be performed to step S315. In step S315, the predicted value to be corrected and the ground information in the predicted value / ground information table T61 are overwritten with the recommended predicted value and the recommended ground information.
[0069]
Next, the exception determination (step S301) processing will be described using the flow shown in FIG.
[0070]
In this flow, the processing from step S401 to step S412 is repeated for each exception condition / recommended value calculation method stored in the exception condition / recommended value calculation method table T81.
[0071]
First, in step S402, the basis information relation 5 (825) is read from the predicted value exception condition 819, the exception threshold value 820, the target basis information 817, the minimum importance 818, and the basis information relation 1 (821). Thereafter, the processing from step S403 to step S411 is repeated for each predicted value / ground information in the predicted value / ground information table T61. In step S404, the predicted value / ground information is read from the predicted value / ground information table T61. Further, the read predicted value / ground information, the product number 613 and the period 614 are the same, and the prediction creating company 611 and the prediction sharing company 612. Read the predicted value / ground information in which is replaced. The pair of these two predicted values / ground information is the target of the exception determination. In step S405, it is checked whether or not the pair of the predicted value and the basis information read in step S404 has been determined to be exceptional under the exceptional condition read in step S402. If it has been determined, the process proceeds to step S411; if not, the process proceeds to step S406. Move on. In step S406, it is determined whether or not the set of predicted values 615 read in step S404 satisfies the predicted value exception condition 819 read in step S402. In step S407, if it is determined in step S406 that the condition is satisfied, the process proceeds to step S410; otherwise, the process proceeds to step S408. In step S408, it is determined whether the set of the basis information 616 read in step S404 applies to the basis information relation 1 (821) to the basis information relation 5 (825), the target basis information 817, and the minimum importance 818 read in step S402. judge. In step S409, if it is determined in step S408 that the condition is satisfied, the process proceeds to step S410; otherwise, the process proceeds to step S411. In step S410, the information of the exception is stored in the exception / recommended predicted value / recommended basis information table T91.
[0072]
Here, the exception / recommended predicted value / recommended basis information table T91 will be described with reference to FIG.
[0073]
The exception / recommended predicted value / recommended basis information table T91 stores one exception and information on the recommended predicted value and the recommended basis information at the time of the exception in one line.
[0074]
As information relating to the exception, an exception number 911 which is an identification number of an exception numbered by the prediction sharing server 11, an exception condition number 912 for identifying an exception condition used for determining the exception, and a creating company 913 which has created the exception condition And its trading company 914 are saved. The product number 915 and the period 916 indicating the product and the period determined as the exception are stored, respectively.
[0075]
Also, as the information on the predicted value and the ground information determined to be an exception, there are a prepared company predicted value 917 and a prepared company ground information 921 which are the predicted value and the ground information of the prepared company 913, and a predicted value and ground information of the trading company 914. The predicted company value 918 and the business company basis information 922 are stored, and the prepared company recommended prediction value 919, the prepared company recommended basis information 923, the business company recommended predicted value 920, and the business company recommended basis information 924 are stored. The created company recommended prediction value 919, the created company recommended basis information 923, the trading company recommended forecast value 920, and the trading company recommended basis information 924 are saved only when the respective values are calculated, and otherwise, nothing is saved. .
[0076]
Also, a correction 925 indicating whether the automatic correction has been performed, a creation company notification 926 indicating whether the exception and the correction of this row are notified to the creation company and the business company, and a business company notification 927 are stored.
[0077]
A method of using the exception / recommended predicted value / recommended basis information table T91 in each process will be described.
[0078]
In step S409, the exception number 911, the exception condition number 912, the creating company 913, the trading company 914, the product number 915, the period 916, the created company predicted value 917, the trading company predicted value 918, the created company basis information 921, and the trading company basis information 922, the creation company notice 926, and the transaction company notice 927 are stored. Further, “uncorrected” is stored in the correction 925. In step S303, the created company recommended predicted value 919 and the transaction company recommended predicted value 920 are stored. In step S304, the created company recommended basis information 923 and the trading company recommended basis information 924 are stored. In step S307, “correction” is stored in the correction 925.
[0079]
Next, the notification process (step S104) will be described using the flow shown in FIG.
[0080]
In this flow, the processing from step S501 to step S512 is repeated for each company stored in the company table T31.
[0081]
First, the company number 311 is read in step S502. After that, the processing from step S503 to step S510 is repeated for each exception, recommended predicted value, and recommended ground information in the exception, recommended predicted value, and recommended ground information table T91.
[0082]
In step S504, the exception, recommended predicted value, and recommended ground information are read from the exception, recommended predicted value, and recommended ground information table T91. Then, it is determined in step S505 whether or not the company number 311 read in step S502 matches the creation company 913. If they match, the process proceeds to step S506, and if they do not match, the process proceeds to step S507. In step S506, it is determined whether or not the creation company notification 926 is “、”. If “作成”, the process proceeds to step S509, and if “x”, the process proceeds to step S510. In step S507, it is determined whether or not the company number 311 read in step S502 matches the trading company 914. If they match, the process proceeds to step S508, and if not, the process proceeds to step S510. In step S508, it is determined whether or not the transaction company notification 927 is “○”. If “○”, the process proceeds to step S509, and if “x”, the process proceeds to step S510.
[0083]
In step S509, the notification screen 41 is created from the exception / recommended predicted value / recommended basis information read in step S504.
[0084]
In step S511, the mail including the notification screen is transmitted to the mail address 312 read from the company table T31 read in step S502.
[0085]
The notification screen 41 displays the following information using all the exception information stored in the exception / recommended predicted value / recommended ground information table T91.
[0086]
First, the exception number 911 is set to the exception number 428, the exception condition number 912 is set to the exception condition 422, the creating company 913 is set to the creating company 423, the trading company 914 is set to the trading company 424, the product number 915 is set to the product number 425, and the period 916 is set. Is displayed in the period 426, and the correction 925 is displayed in the correction 427. The correction 427 displays “completed” when the correction 925 is “correction”, and displays “not yet” when the correction 925 is “uncorrected”.
[0087]
Next, when the correction 925 is “correction”, the created company prediction value 917 and the transaction company prediction value 918 are set to the pre-correction prediction value 4211, and the created company basis information 921 and the transaction company basis information 922 are set to the pre-correction basis information 4212. The created company recommended forecast value 919 and the trading company recommended forecast value 920 are displayed in the corrected forecast value 4213, and the created company recommended ground information 923 and the trading company recommended ground information 924 are displayed in the corrected ground information 4214. The link 429 displays a more detailed link to the exception screen 51.
[0088]
When the correction 925 is “uncorrected”, the created company forecast value 917 and the business company forecast value 918 are used as the forecast value 4311, the created company basis information 921 and the business company basis information 922 are used as the basis information 4312, and the created company recommended forecast value 919. And the recommended company value 920 are displayed in the recommended value 4313, and the recommended company information 923 and the recommended company information 924 are displayed in the recommended information 4314.
[0089]
For example, the row 9101 is displayed as the exception 421, the row 9102 is displayed as the exception 431, and the row 9103 is displayed as the exception 441. The link 429 displays a more detailed link to the exception screen 51.
[0090]
The sales company terminal 12 or the purchase company terminal 13 that has received the notification screen 41 can see the more detailed exception screen 51 by clicking the link 429.
[0091]
The exception screen 51 shown in FIG. 5 includes the exception / recommended predicted value / recommended exception / recommended predicted value / recommended basis information of the line of the exception number 911 corresponding to the exception number 428 corresponding to the link 429 clicked on the notification screen 41. It reads from the ground information table T91 and further reads the predicted value / ground information from the predicted value / ground information table T61, and displays the following information.
[0092]
First, the creating company 913 is displayed as the creating company 521, the trading company 914 is displayed as the trading company 522, and the product number 915 and the product number 523 are displayed.
[0093]
Also, the predicted value 615 of the row of the prediction creating company 611 that matches the created company 521 and the trading company 522 in the created company forecast value 503 and the trading company forecast value 504, respectively, is included in the created company basis information 505 and the trading company basis information 506. The basis information 616 of the row of the forecast creation company 611 that matches the creation company 521 and the transaction company 522, respectively, is displayed in chronological order for each period 502.
[0094]
Further, the following is displayed in a column 5001 of the period 502 corresponding to the period 916. When the correction 925 is “correction”, the exception number 911 is displayed in the exception number 517, the prepared company forecast value 917 is the pre-preparation company correction value 511, the transaction company prediction value 918 is the transaction company pre-transaction value 512, the basis of the preparation company. The information 921 is displayed in the base information before creation company correction 513, and the business base information 922 is displayed in the base information before business company correction 514. Further, “automatic correction” is displayed in the status 516, and the list 541 is displayed in the operation 516. Enter a correction value in the fields of the created company forecast value 503, the business company forecast value 504, the created company basis information 505, and the business company basis information 506, select "correct to input value" from the list 541, and press the enter button 531. Is corrected to the input value, and selecting “undo the correction” from the list 541, the predicted value 615 and the ground information 616 of the corresponding row in the predicted value / ground information table T61 are changed from the value 512 before the creation company correction. It is overwritten using the trading company pre-correction basis information 514.
[0095]
When the correction 925 is “uncorrected”, the exception number 911 is displayed in the exception number 517 as shown in the exception screen 52 of FIG. 14, the created company recommended forecast value 919 is created, the created company recommended value 508, and the trading company recommended forecast is displayed. The value 920 is stored in the recommended company value 509, the created company recommended basis information 923 is stored in the created company recommended basis information 510, and the recommended company basis information 924 is stored in the recommended company basis information 511. “Exception” is displayed in the status 516, and the list 542 is displayed in the operation 516. When a correction value is input in the fields of the created company forecast value 503, the business company forecast value 504, the created company basis information 505, and the business company basis information 506, the user selects "correct to the input value" from the list 542 and presses the enter button 531. When the predicted value 615 and the ground information 616 in the corresponding row of the predicted value / ground information table T61 are corrected to the input values, and "correction to recommended value" is selected from the list 542, the correspondence of the predicted value / ground information table T61 is obtained. The predicted value 615 and the basis information 616 of the row to be executed are overwritten from the created company recommended value 508 with the transaction company recommended basis information 510.
[0096]
For example, when the link 429 of the notification screen 41 is clicked, the exception screen 51 is displayed. When the link 439 is clicked, the exception screen 52 shown in FIG. 14 is displayed. When the link 449 is clicked. Displays an exception screen 53 shown in FIG. Here, one exception, recommended predicted value, and recommended ground information are displayed on one screen, but a plurality of exception, recommended predicted value, and recommended ground information may be displayed on one screen.
[0097]
The detailed embodiment has been described above. Hereinafter, a specific joint prediction method will be described with reference to FIGS. 6 to 9.
[0098]
The basis information definition of the rows 7101 to 7112 in FIG. 7 is set in advance by the selling company (company number: 1001) and the purchasing company (company number: 1002). Exceptional conditions and recommended value calculation methods in lines 8101 to 8106 are set. Then, the predicted values and the basis information of the rows 6101 to 6112 shown in FIG. 6 are registered from the selling company (company number: 1001), and the predicted values and the basis information of the rows 6201 to 6212 from the purchasing company (company number: 1002). Is registered.
[0099]
The prediction sharing server 11 first performs an exception determination in the following procedure.
[0100]
The exception condition stored in the row 8101 is read. From the creating company 812 and the trading company 813, the product number 814, the period start 815, and the period ending 816 of the exceptional condition, the forecast creating company 611 in FIG. 6 is a selling company (company number: 1001), and the forecast sharing company 612 is a purchasing company ( The company number: 1002), the product number 613 is 0001, and the period from June 1 to June 4 is the forecast value, the basis information, and the forecast creation company 611 is the purchasing company (company number: 1002) and the forecast sharing company 612 is the A sales company (company number: 1001), a product number 613 is 0001, and a period from June 1st to June 4th week is a predicted value. It becomes the basis information. According to this application condition, the set of the row 6101 and the row 6201, the set of the row 6102 and the row 6202, the set of the row 6103 and the row 6203, and the set of the row 6104 and the row 6204 are the predicted values and the basis information to which the exceptional condition of the row 8101 is applied. Is detected as
[0101]
First, an exception determination is performed by applying the predicted value exception condition and the basis information exception condition to the pair of the row 6101 and the row 6201. The exception condition 1 for the prediction value set in the prediction value exception condition 819 is applied to the prediction value 100 in the row 6101 and the prediction value 95 in the row 6201. Since the exception condition 1 is the radio button 251 “no exception judgment” in FIG. 2, it is not judged as an exception. In addition, the basis information exception condition is the basis information relation 3 (823) from the basis information relation 1 (821) to the basis information relation 5 (825), that is, the check box 263 in FIG. Part of (but does not match). " Neither the basis information in the row 6101 nor the row 6201 exists and does not match the basis information relation 3 (823), so that it is not determined to be an exception. As described above, since neither the predicted value nor the ground information is determined to be an exception, the rows 6101 and 6201 are not determined to be an exception.
[0102]
Next, the predicted value exception condition and the ground information exception condition are applied to the pair of the row 6102 and the row 6202 to perform an exception determination. The exception condition 7 for the predicted value set in the predicted value exception condition 819 is applied to the predicted value 100 in the line 6102 and the predicted value 206 in the line 6201. The exception condition 7 is a radio button 257 “| Predicted value of created company−Predicted value of trading company | ≧ exception threshold” in FIG. 2 and the exception threshold 820 is 30, so the above prediction value matches the condition, and the exception judge. Since the exception is determined, the information of the exception is stored as shown in a row 9101 in FIG. At this point, the columns 919, 920, 923, and 924 are not entered because the recommended value has not been calculated yet.
[0103]
Similarly, even if an exception is determined for a set of the row 6103 and the row 6203 and a set of the row 6104 and the row 6204, no exception is determined.
[0104]
As described above, the exception determination based on the exception condition in the row 8101 ends.
[0105]
Similarly, when an exception is determined based on the exceptional conditions in lines 8202 to 8206, an exception is determined in the pair of lines 6105 and 6205 according to the exceptional condition in line 8204, and the exception is determined in the pair of lines 6109 and 6209 according to the exceptional condition in line 8205. It is determined that there is an exception, and information of each exception is stored in rows 9102 and 9103 of FIG.
[0106]
An exception determination process for these two exceptions will be described.
[0107]
Since the predicted value exception condition 819 in each of the rows 8204 and 8205 is “no exception occurred”, the predicted values of the set of the row 6105 and the row 6205 and the set of the row 6109 and the row 6209 are not determined to be exceptions.
[0108]
The basis information in the rows 6105 and 6205 is the basis information K1 (importance level 2) and the basis information K4 (importance level 4). By applying the target ground information 816 of the row 8204 to this, that is, the “most important ground information”, only the ground information K4 having the importance level of 4 is an exception determination target. According to the basis information relation 1 (821) to the basis information relation 5 (825) in the row 8204, the exception condition for the basis information is 3, that is, the check box 263 in FIG. (They do not match) ", and there is no target company information of the creating company, ie, row 6105, and the target company information of the transaction company, ie, row 6205, is the basis information K4. I do.
[0109]
The basis information in lines 6109 and 6209 is basis information K2 (magazine feature, importance 4) and basis information K5 (sale 50% OFF, importance 4). By applying the target ground information 816 of the row 8205, that is, the “most important ground information” to this, the ground information K2 and the ground information K4 having the importance level of 4 are subject to exception determination. From the basis information relation 1 (821) to the basis information relation 5 (825) in the row 8204, the exception conditions for the basis information are 4 and 5, that is, the check box 264 in FIG. It is determined that an exception occurs when one of the following conditions is satisfied: “the base information of the creating company does not match the base information of the transaction company at all”. The creator company, that is, the target rationale information of row 6109 is the rationale information K2, and the trading company, that is, the rationale information of row 6209 is the rationale information K5. Since it is applicable, it is determined as an exception.
[0110]
Next, the prediction sharing server 11 creates and automatically corrects recommended values for the exceptions stored in the rows 9101 to 9103 in FIG. 9 according to the following procedures (1) to (3).
(1) Recommended value creation and automatic correction for the exception in line 9101
From the exception condition number 912 (exception condition number: 1) in the row 9101, a row 8101 which is the exception condition (exception condition number: 1) of this exception is read. The recommended prediction value calculation method 826 is 3, that is, the radio button 273 in FIG. 2 "match the predicted value of the creating company with the predicted value of the trading company". By applying to the predicted value 918 (predicted value: 206), the recommended predicted value 919 of the creating company is calculated as 206.
[0111]
The target basis information 817 in the row 8101 is 3, that is, “the most important basis information” from the radio button 234 in FIG. 2, and the recommended basis information calculation method 827 is 3, ie, the radio button 283 in FIG. According to the information of the trading company. " These are applied to the created company basis information 921 (ground information: none) and the predicted value of the trading company 918 (ground information: 2, importance: 2), and the recommended basis information 923 of the created company is calculated as 2.
[0112]
Next, since the automatic correction 828 in the row 8101 is “O”, the predicted value 615 of the row 6102 that matches the creation company 913, the transaction company 914, the product number 915, and the period 916 of the row 9101 is changed to the created company recommended prediction value 919. (Predicted value: 206).
(2) Recommended value creation and automatic correction for the exception in line 9102
From the exception condition number 912 (exception condition number: 4) of the row 9102, a row 8104 which is the exception condition (exception condition number: 4) of this exception is read. The recommended prediction value calculation method 826 is 3, that is, the radio button 273 in FIG. 2 “Match the predicted value of the creating company to the predicted value of the trading company”. By applying to the predicted value 918 (predicted value: 332), the recommended predicted value 919 of the creating company is calculated as 332.
[0113]
The target basis information 817 of the row 8104 is 3, that is, “the most important basis information” from the radio button 234 of FIG. 2, and the recommended basis information calculation method 827 is 3, that is, the radio button 283 of FIG. According to the information of the trading company. " These are applied to the created company basis information 921 (ground information: 2, importance: 2) and the predicted value of the trading company 918 (base information: 4, importance: 4), and the recommended basis information 923 of the created company is set to 4. calculate.
[0114]
Since the automatic correction 828 in the row 8104 is “x”, the automatic correction is not executed.
(3) Creation and automatic correction of recommended values for the exception in line 9103
From the exception condition number 912 (exception condition number: 5) in the row 9103, a row 8105 which is the exception condition (exception condition number: 5) of this exception is read. The recommended prediction value calculation method 826 is 1, that is, the radio button 271 in FIG. 2 is “do not calculate”, and the recommended prediction value is not calculated.
The target basis information 817 of the row 8105 is 3, that is, “the most important basis information” from the radio button 234 of FIG. 2, and the recommended basis information calculation method 827 is 5, that is, the radio button 285 of FIG. Add all the evidence information in at least one of them. " These are applied to the created company base information 921 (base information: 11, importance level: 4) and the predicted value of the trading company 918 (base information: 4, importance level: 4). The base information 924 is calculated as 4 or 12.
[0115]
Since the automatic correction 828 in the row 8105 is “x”, the automatic correction is not executed.
[0116]
Finally, the prediction sharing server 11 notifies the sales company terminal 12 and the sales company terminal 13 of the exception, the recommended predicted value, and the recommended basis information stored in the rows 9101 to 9103 of FIG. 9 in the following procedure.
[0117]
First, the company number 311 (company number: 1001) is read from the row 3101 of the company table T31 in FIG. 13, and a row in which either the creation company 913 or the transaction company 914 in the rows 9101 to 9103 matches is searched. All creation companies 913 in the rows 9101 to 9103 are 1001 and match the company number 311 in the row 3101. Therefore, the creation company notification 926 in the row 9101 is read. Since the value is “○”, the exception 421 in FIG. 10 is created using the columns 911 to 925. Similarly, since the creation company notification 926 in the rows 9102 and 9103 is also “○”, exceptions 422 and 423 in FIG. 10 are created, respectively. The notification screen 41 created as described above is transmitted to the mail address 312 (mail address: mail@hanbai.co.jp) in the row 3101.
[0118]
Next, the company number 311 (company number: 1002) is read from the row 3102 of the company table T31 in FIG. 13, and a row in which either the creation company 913 or the transaction company 914 in the rows 9101 to 9103 matches is searched. All the trading companies 914 in the rows 9101 to 9103 are 1002, which matches the company number 311 in the row 3102. Therefore, the transaction company notification 927 in the row 9101 is read. Since the value is “○”, the exception 471 in FIG. 11 is created using the columns 911 to 925. Since the creation company notification 926 in the rows 9102 and 9103 is “×”, the exception information is not displayed on the notification screen. The notification screen 45 created as described above is transmitted to the mail address 312 (mail address: mail@koubai.co.jp) in the row 3102.
[0119]
When the link 414 or the link 464 is clicked by the sales company terminal 12 or the purchase company terminal 13, the shared company server 11 transmits the predicted value / ground information table T61 shown in FIG. 6 and the exception / recommended predicted value / recommended shown in FIG. The exception screen 51 shown in FIG. 5 is displayed using the information of the basis information table T91.
[0120]
As described above, according to the present embodiment, ground information is added to the predicted value, and an exception to the ground information is determined. Differences of opinion can be found by exception judgment. When it is determined that an exception has occurred, the cause of the exception can be easily specified by displaying the basis information. Furthermore, by using the basis information of both at the time of the exception, the recommended value of the predicted value correction can be automatically calculated, and the predicted value can be easily corrected.
[0121]
【The invention's effect】
According to the present invention, it is possible to provide an exception judgment that can not be found because the predicted values coincide with each other by chance, and that a difference in views on future demand can be found, and it is possible to easily identify an exception factor when it is judged as an exception. This has a remarkable effect that the predicted value can be easily corrected.
[Brief description of the drawings]
FIG. 1 is a flowchart showing a system configuration and a processing flow in an embodiment of the present invention.
FIG. 2 is an exception condition / recommended value calculation method setting screen according to the embodiment of the present invention.
FIG. 3 is a flowchart of an exception determination process, a recommended value calculation process, and an automatic correction process according to the embodiment of the present invention.
FIG. 4 is a flowchart of an exception determination process according to the embodiment of the present invention.
FIG. 5 is an exception screen according to the embodiment of the present invention.
FIG. 6 is a predicted value / ground information table according to the embodiment of the present invention.
FIG. 7 is a ground information definition table in the embodiment of the present invention.
FIG. 8 is an exception condition / recommended value calculation method table in the embodiment of the present invention.
FIG. 9 is an exception / recommended predicted value / recommended basis information table in the embodiment of the present invention.
FIG. 10 is a screen for notifying a sales company according to the embodiment of the present invention.
FIG. 11 is a notification screen to a purchasing company in the embodiment of the present invention.
FIG. 12 is a flowchart of a notification process according to the embodiment of the present invention.
FIG. 13 is a company table according to the embodiment of the present invention.
FIG. 14 is an exception screen according to the embodiment of the present invention.
FIG. 15 is an exception screen according to the embodiment of the present invention.
[Explanation of symbols]
11: prediction sharing server, 12: selling company terminal, 13: purchasing company terminal, 14: communication network.

Claims (9)

In a computer-based predictive value evaluation method for evaluating predictive values created by a plurality of companies, a predictive value that is a condition for determining a predictive value exception that is a state of a difference between the predictive values and a predictive value created by each company Applying the exceptional condition, the predicted value exception determination processing for determining the predicted value exception, and the ground information which is a state of the difference of the ground information to the ground information which is a factor for calculating the predicted value created by each company. Applying one or both of the basis information exception judgment processes to determine the basis information exception by applying the basis information exception condition that is the condition for judging the exception, and the state of the difference between the predictions created by multiple companies A predictive value evaluation method characterized by determining an exception. 2. The prediction value evaluation method according to claim 1, further comprising adding importance to the ground information, and specifying ground information by the importance in the ground information exception determination process. 3. The prediction value evaluation according to claim 1, further comprising, if the exception is determined, executing a recommended prediction value calculation process for calculating a recommended prediction value for correcting the prediction value. Method. 4. The prediction value evaluation according to claim 1, further comprising, when it is determined that the exception is present, executing recommended ground information calculation processing for calculating recommended ground information for correcting the ground information. Method. 5. The predictive value evaluation method according to claim 4, wherein importance is further added to the ground information, and ground information to be subjected to the recommended ground information calculation processing is designated by the importance. The prediction value evaluation method according to claim 3, further comprising correcting the predicted value or the ground information based on the recommended predicted value or the recommended ground information. Predicted value exception determination processing for determining a predicted value exception by applying a predicted value exception condition that is a condition for determining a predicted value exception that is a state of a difference between predicted values to a predicted value created by each of a plurality of companies. And a basis information exception condition that is a condition for judging a basis information exception that is a state of a basis information difference is applied to basis information that is a factor for calculating a predicted value created by each company, Is a program for executing one or both of the grounds information exception determination processing to determine the difference between predictions created by a plurality of companies, and causing the computer to realize a function of finding a difference between the predictions. 8. The program according to claim 7, further comprising the step of, when the exception is determined, causing the computer to execute a recommended predicted value calculation process for calculating a recommended predicted value for correcting the predicted value. 9. The computer-readable storage medium according to claim 7, wherein the computer further executes a recommended basis information calculation process for calculating recommended basis information for correcting the basis information when the exception is determined.

Images